Study on Feature Extraction of Engine Vibration Signal Based on Time-frequency Singular value spectrum

PDF(1189 KB)

Journal of Vibration and Shock ›› 2010, Vol. 29 ›› Issue (9) : 222-225.

论文

Study on Feature Extraction of Engine Vibration Signal Based on Time-frequency Singular value spectrum

Wu Dinghai;Zhang Peilin;Zhang Yingtang;Xu Chao

Author information +

History +

Abstract

A fault feature extraction method of engine vibration signal based on Time-frequency distribution and Singular value spectrum was proposed. The time-frequency distribution based on Complex Morlet wavelet transform can describe both the local feature and energy distribution. With Singular value decomposition, the time-frequency distribution matrix would be decomposed as singular value series which can forthright describe the basic time-frequency characters. With the feature extraction and feature selection based on separability evaluation, the time-frequency information was dimensionality reduction and the efficiency features were extracted. The test results illustrate that the features extracted by the mentioned techniques have the capacity of anti-noise-interference and enhanced sort separability.

Key words

Time-frequency distribution / Singular value decomposition / Complex Morlet wavelet / Feature extraction

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Wu Dinghai;Zhang Peilin;Zhang Yingtang;Xu Chao. Study on Feature Extraction of Engine Vibration Signal Based on Time-frequency Singular value spectrum [J]. Journal of Vibration and Shock, 2010, 29(9): 222-225